This research was aimed to investigate the enhancement of electrokinetic (EK) remediation arsenate-contaminated soil by permeable reaction barrier (PRB). All experiments, which experimental parameters included the position, materials, and quantity of PRB, processing fluid types, potential gradients, and treatment time, were conducted in two types of EK systems. One was Pyrex glass cylindrical cells with dimension of 4.2 cm (£r) ¡Ñ 12 cm (L) and the other was a small pilot-scale modulus with dimension of 36cm (L) ¡Ñ18cm (W) ¡Ñ18cm cm (H). The PRBs were composed of four kinds of reaction materials, which included commercial zero valent iron (Fe(0)C), manufactured zero valent iron (Fe(0)M), commercial hydrous ferric oxide (FeOOHC), and manufactured hydrous ferric oxide (FeOOHM), mixed with ottawa sand in a ratio of 1:2,respectively, and installed in the anode, middle, and cathode side of the EK systems. For 5-day EK cylindrical cell tests, the results showed that the PRB installation would result in a lower electroosmosis permeability (Ke) and a higher removal efficiency of arsenate. The arsenate removal efficiency of EK system with PRB was in the range of 43.89-70.25%, which was 1.5~2.6 times greater than that without PRB, and the value of Ke was in the range of 4.30-12.61¡Ñ10-6 cm2/V-s. The soil pH after EK/PRB treatment was much closer to natural and more arsenate was collected in the anode reservoir. Moreover, the remediation performance of FeOOHC as PRB materials was much better than other materials. For EK pilot-scale modulus tests, it was shown that the removal efficiency of arsenate was effectively enhanced as improved experimental parameters and, however, led to increase the treatment cost. In EK modulus without PRB, the removal efficiency of arsenate, elctroosmosis permeability, and energy consumption were 27.76%, 3.30-5.39¡Ñ10-6 cm2/V-s, and 1724.81 kWh/m3, respectively. Furthermore, the treatment cost was NT 9583/m3. As increasing treatment time, graphite electrode, potential gradient, and quantity of PRB materials, the removal efficiency of arsenate increased to as high as 45.11-71.22% and the treatment cost also increased up to NT 24,800-57,730/m3. As investigated the binding form of arsenate with soil after EK/PRB treatment, it was found that the arsenate ¡Vsoil binding forms of Fe-Mn oxide bound, organically bound, and residual in the soil section behind the PRB were much easier transformed to the forms of exchangeable and carbonate bound. The transformation rate reached as high as 72.5% and it increased with treatment time. However, the Fe-Mn oxide bound was still the main binding form, 61.6-81.6%, in the soil section prior to the PRB. The removal mechanism of arsenate contaminated soil remediation was dominated by electromigration, electrolysis, and electroosmosis in EK system without PRB. And, in EK/PRB system, the removal of arsenate from soil was mainly resulted from adsorption rather than redox reaction by PRB. To sum up, the PRB can effectively enhance the electrokinetic remediation of arsenate contaminated soil by choosing the right PRB materials and operation parameters.